Pflügers Archiv

, Volume 451, Issue 3, pp 474–478

Fenamates and diltiazem modulate lipid-sensitive mechano-gated 2P domain K+ channels

Authors

    • Department of Visual Neuroscience & OphthalmologyKanazawa University
  • Mayumi Sakurai
    • Department of Visual Neuroscience & OphthalmologyKanazawa University
  • Norimasa Sakurada
    • Department of Visual Neuroscience & OphthalmologyKanazawa University
  • Kazuhisa Sugiyama
    • Department of Visual Neuroscience & OphthalmologyKanazawa University
Ion Channels, Transporters

DOI: 10.1007/s00424-005-1492-5

Cite this article as:
Takahira, M., Sakurai, M., Sakurada, N. et al. Pflugers Arch - Eur J Physiol (2005) 451: 474. doi:10.1007/s00424-005-1492-5

Abstract

A swelling-activated, background K+ current in the corneal epithelium is characteristically activated by fenamates and inhibited by diltiazem. Fatty acids also stimulate this current, indicating that its origin is a lipid-sensitive mechano-gated 2P domain K+ channel. In the present study, modulation of TREK-1, TREK-2, and TRAAK channels by fenamates and diltiazem was examined. TREK-1, TREK-2, and TRAAK currents transiently expressed in COS-7 cells were recorded by the perforated-patch configuration. As previously reported, arachidonic acid (20 μM) stimulated all of these channels, and a volatile anesthetic, halothane (1 mM) augmented TREK-1 and TREK-2 but not TRAAK. Flufenamic acid (FA, 100 μM), niflumic acid (NA, 100 μM), and mefenamic acid (MA, 100 μM) markedly stimulated TREK-1, TREK-2, and TRAAK. The potency sequence for the activation of TREK-1 and TREK-2 was FA > NA = MA, and the potency sequence for the activation of TRAAK was FA = NA > MA. Diltiazem (1 mM) inhibited TREK-1 and TREK-2, but not TRAAK. In conclusion, fenamates are openers of the lipid-sensitive mechano-gated 2P domain K+ channels, and diltiazem may be a specific blocker for TREK. These novel findings could help to further understand channel functions of the mechano-gated 2P domain K+ channels.

Keywords

FenamatesTREKTRAAKDiltiazemPotassium channels2P domainCorneal epithelium

Introduction

2P domain K+ channels (designated KCNK by the Human Genome Organization) are defined by their common structure consisting of four transmembrane domains and two pore domains. They are classified into several functional groups, one of which is the lipid-sensitive mechano-gated 2P domain K+ channels, comprising TREK-1 (TWIK-related K+ channel, TWIK; Tandem of pore domains in a weak, inward rectifying K+ channel) (KCNK2) [5, 6], TREK-2 (KCNK10) [2, 6], and TWIK-related arachidonic acid-stimulated K+ channel (TRAAK) (KCNK4) [3, 6]. These channels are widespread in the central nervous system and are reported to participate in neuroprotection and in general anesthesia. To date, several openers of the mechano-gated 2P domain K+ channels have been identified: polyunsaturated fatty acids including arachidonic acid, phospholipids (for all of TREK-1, TREK-2, and TRAAK), intracellular acidosis and volatile general anesthetics (for TREKs only) [6].

The corneal epithelium serves as a transparent refractive surface and ocular barrier, playing crucial roles for vision. A mechano-gated K+ channel, namely the large conductance K+ channel is responsible for the major, background K+ conductance in the corneal epithelium, which is characteristically stimulated by fenamates and inhibited by diltiazem [7]. We recently discovered that this K+ current in the bovine corneal epithelium is also enhanced by certain unsaturated fatty acids, including arachidonic acid [8]. Electrophysiological and pharmacological properties of this corneal K+ channel indicate that it may be one of the lipid-sensitive mechano-gated 2P domain K+ channels. At the same time, it raises the question as to whether these 2P domain K+ channels are stimulated by fenamates and inhibited by diltiazem.

In the present study, we investigated novel pharmacological properties of the lipid-sensitive mechano-gated 2P domain K+ channels, i.e., effects of fenamates and diltiazem on TREK-1, TREK-2, and TRAAK currents transiently expressed in COS-7 cells.

Methods

Transient expression of 2 pore domain potassium channels in COS-7

For cloning TREK-1 and TREK-2 mRNAs, cDNA was synthesized from the polyA+ RNA of human corneal epithelial cells (HCEC, strain#9C0805,#0C0617,#0C0806, Kurabo Industries, Osaka, Japan) cultured in EpiLife Medium with Human Corneal Growth Supplement (Kurabo Industries). Each coding region (open reading frame) was amplified by TAKARA Ex Taq. The coding region of TRAAK was amplified by PCR from human brain cDNA (BD Biosciences Clontech, CA, USA). The primers were synthesized based on the sequence information (Genbank accession number AF129399 of TREK-1, AF279890 of TREK-2, and NM01611 of TRAAK) [2, 3, 5]. The nucleotide sequences of the primers are available upon request. In the deduced amino acid sequences, we found two substitutions: at 376 position of asparagine to serine in TREK-1, and at 354 position of leucine to proline in TRAAK. Among splicing variants for TREK-2, the cloned products was identical to TREK-2a. The amplified cDNAs were inserted into the multi-cloning site of an expression vector pIRES2-EGFP (BD Biosciences Clontech). COS-7 cells were cultured in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum. Each plasmid was transfected into COS-7 cells using Lipofectamine 2000 (Invitrogen Corporation, Carlsbad, CA, USA). The positive cells were visualized by fluorescent microscope at 24 h after transfection. The procedure of gene recombination was approved by the institutional ethical review committee.

Solutions

External and internal solutions were essentially the same as those used in our previous patch-clamp study [8]. Briefly, the standard extracellular solution consisted of (in mM): 135 NaCl, 5.0 KCl, 10 HEPES, 10 glucose, 1.8 CaCl2 and 1.0 MgCl2, and was titrated to pH 7.4 with NaOH. The standard pipette solution consisted of (in mM unless noted otherwise): 30 KCl, 83 potassium gluconate, 5.0 HEPES, 5.5 EGTA-KOH, 0.5 CaCl2 and 2.0 MgCl2, amphotericin B (240 μg/ml), and was titrated to pH 7.2 with KOH.

Perfusates containing fenamates and arachidonic acid were prepared just before the experiments by diluting stock solutions in DMSO. Since final concentrations of DMSO in perfusates never exceeded 0.2% throughout the experiments, the presence of DMSO in perfusates did not affect the whole cell currents in this study.

NaCl, KCl, CaCl2 were obtained from Wako Chemical Co. (Osaka, Japan). All other reagents were obtained from Sigma Chemical Co. (St. Louis, MO, USA).

Electrophysiological methods

Cell suspensions were transferred to a Lucite perfusion chamber (RC-5/25, Warner Instrument, Hamden, CT) and perfused at a flow rate of 0.7 ml/min by a perfusion system (BPS-4, ALA Scientific Instruments, Westbury, NY, USA). All experiments were conducted at room temperature (20–25°C).

Patch pipettes were pulled from borosilicate glass tubing (BF150-110-10, Sutter Instruments, San Rafael, CA, USA) with a multistage programmable puller (P-97, Sutter Instruments, San Rafael, CA, USA). The pipette input resistance was between 1 and 3 MΩ. Gigaohm seal was established under phase-contrast microscopy (ECLIPSE TE300, Nikon, Tokyo, Japan), using a micromanipulator (MP-285, Sutter Instruments) Currents under voltage-clamp were recorded by a Heka amplifier system (EPC-8, Heka, Lambrecht/Pfalz, Germany). The built-in low pass filter was set to 3 kHz. Recordings were referenced to an Ag–AgCl electrode (EP-2, WPI, Sarasota, FL, USA). The membrane capacitance was compensated by built-in circuits. The apparent membrane potential was corrected by the pipette tip potential (10 mV). Statistical data are presented as mean ± SD.

Results

TREK and TRAAK currents transiently expressed in COS cells

COS cells transfected with TRAAK, TREK-1 or TREK-2 cDNA were visualized by their fluorescence. In COS cells expressing TREK-1, TREK-2, and TRAAK, the whole cell current was dominated by an outwardly rectifying, noisy current (Figs. 1b, 2a). Just after the gigaohm sealing of cells expressing TREK or TRAAK, gentle suctioning the pipette to stretch the cell membrane caused marked activation of noisy, outward currents (not shown), indicating mechano-sensitivity of the channels. Similar currents were never seen in the control cells transfected with pIRES2-EGFP vector only (n=5, not shown). Amplitudes of TREK and TRAAK currents varied considerably among cells: an outward current at +40 mV (I+40) was 1.2±1.0 nA in TREK-1 (n=41), 2.2±1.3 nA in TREK-2 (n=57), and 1.1±0.7 nA in TRAAK (n=29). The zero-current potentials of the cells were −61±10 mV in TREK-1 (n=41), −67±8 mV in TREK-2 (n=57), and −66±7 mV in TRAAK (n=29). As summarized in Fig. 1a, 20 μM arachidonic acid augmented I+40 of TREK-1, TREK-2, and TRAAK to 200–300% of the control value, a finding consistent with other reports [6]. Halothane (1 mM), a volatile anesthetic known as an opener of TREK channels [5, 6], increased I+40 of TREK-1 and TREK-2, but did not significantly alter I+40 of TRAAK (Fig. 1a). On the basis of these electrophysiological and pharmacological properties consistent with previous reports [6], we confirmed that TREK and TRAAK channels were expressed in COS-7 cells.
Fig. 1

ad Stimulation of TREK-1, TREK-2, and TRAAK. a Relative current amplitudes at +40 mV (I+40) in TREK-1, TREK-2, and TRAAK by application of 20 μM arachidonic acid (open bar, n=9, 10, 6, respectively) and 1 mM halothane (closed bar, n=3, 5, 8, respectively) were averaged. Significant changes by paired student t-test are shown by asterisks (**: P<0.01, *: P<0.05). b In COS-7 cells expressing TREK-1 (left), TREK-2 (middle), and TRAAK (right), whole cell currents elicited by voltage steps ranging from + 50 to −60 mV (10 mV increments) from a holding potential of −10 mV were measured in the absence and presence of 100 μM flufenamic acid. c Dose dependence of channel activation by flufenamic acid. The fractional responses to flufenamic acid were measured in TREK-1 (n=4), TREK-2 (n=5), and TRAAK (n=4). The maximum current magnitude in each cell was obtained in the presence of 1 mM flufenamic acid. d Relative I+40 in TREK-1, TREK-2, and TRAAK by application of 100 μM flufenamic acid (open bar, n=6, 7, 6, respectively), 100 μM niflumic acid (closed bar, n=9, 6, 7, respectively), and 100 μM mefenamic acid (gray bar, n=10, 5, 5, respectively) were averaged. Significant differences by student t-test (one-tail) were shown by asterisks (**: P<0.01, *: P<0.05). All error bars in this figure represent SD

Fig. 2

ac Inhibitory effects of diltiazem on 2P domain K+ channels. a In representative COS-7 cells expressing TREK-1, TREK-2, and TRAAK, whole cell currents elicited by voltage steps ranging from + 50 to −60 mV were obtained in the absence and presence of 1 mM diltiazem. b Relative current amplitude at +40 mV (I+40) by application of 1 mM diltiazem was averaged in TREK-1 (n=7), TREK-2 (n=5), and TRAAK (n=7). Significant changes in paired student t-test were shown by asterisks (**: P<0.01). c Dose response curves. Relative current amplitude at each concentration was measured for TREK-1 (n=6) and TREK-2 (n=5). Smooth curves represent fitting the data to the first order equation: Idiltiazem/Icontrol = 1− [diltiazem]/([diltiazem]+IC50), with IC50 of 1.8×10−4 M for TREK-1 and 3.3×10−4 M for TREK-2. All error bars represent SD

Sensitivities to fenamates

To investigate the effects of fenamates on the lipid-sensitive mechano-gated 2P domain K+ channels, we used three kinds of fenamates-flufenamic acid (FA), niflumic acid (NA), and mefenamic acid (MA)—which have commonly been used in previous studies on Ca2+-activated potassium channels stimulated by fenamates [1, 4]. In COS cells expressing TREK-1, TREK-2, and TRAAK, FA (100 μM) significantly augmented K+ current (Fig. 1b). The current components augmented by FA were similar to those augmented by arachidonic acid: noisy, outwardly rectifying, with a reversal potential near EK (−80 mV). Since FA above 1 mM usually activated leak currents, the maximum responses of the channels were obtained in the presence of 1 mM FA. Stimulatory effects of FA on these channels showed similar dose-dependence, with the estimated EC50 close to 100 μM (Fig. 1c). The other two fenamates tested, NA and MA, also enhanced TREK-1, TREK-2, and TRAAK with similar dose-dependence (n=4 in each channel, not shown). In each channels, the activation potencies by these fenamates (100 μM) were compared (Fig. 1d): the potency sequence for the activation of TREK-1 and TREK-2 was FA > NA = MA and the potency sequence for the activation of TRAAK was FA = NA> MA.

Sensitivities to diltiazem

Sensitivities of TREK and TRAAK to diltiazem, a specific inhibitor of the large conductance K+ current in corneal epithelium [7, 8] were examined. One millimolar diltiazem significantly inhibited TREK-1 and TREK-2, but not TRAAK expressed in the COS cells (Fig. 2a, b). Milimolars of diltiazem inhibited TREK-1 and TREK-2 almost completely, and IC50 of diltiazem was estimated to be 1.8×10−4 M for TREK-1 and 3.3×10−4 M for TREK-2 (Fig. 2c).

Discussion

With regard to channel activation of the lipid-sensitive mechano-gated 2P domain K+ channels, the following pharmacological properties have been elucidated: (1) TREKs and TRAAK are activated by polyunsaturated fatty acids, including arachidonic acid and phospholipids, (2) TREKs are activated by intracellular acidosis, (3) TREKs are activated by volatile general anesthetics, e.g., halothane [2, 3, 5, 6]. These functional characteristics enable the lipid-sensitive mechano-gated 2P domain K+ channels to play critical roles in neuroprotection of the central nervous system and in general anesthesia, and in the maintenance of cell homeostasis in peripheral organs. To date, the effects of fenamates on 2P domain K+ channels have never been mentioned in published reports. Fenamates, a family of nonsteroidal anti-inflammatory drugs (NSAIDs), are classically known as inhibitors of anion permeability, especially chloride channels. Fenamates also inhibit other ionic permeability, including gap junctions, nonselective cation channels, NMDA receptors, and a voltage-gated K+ channel, Kv2.1. On the other hand, stimulatory effects of fenamates have been reported only for K+ channels, including Ca2+-activated K+ channels [1, 4]. The stimulatory potencies of flufenamic and niflumic acids on Ca2+-activated K+ channels are reported to be higher than the stimulatory potency of mefenamic acid [1, 4], a potency sequence similar to the results of the present study for the 2P domain K+ channels (Fig. 1d).

The hypothesis that diltiazem may inhibit the lipid-sensitive mechano-gated 2P domain K+ channels arose from studies of a lipid-sensitive K+ channel in the corneal epithelium, that is characteristically inhibited by diltiazem [7, 8]. Sensitivities of TREKs and TRAAK to channel blockers are somewhat unique: these are potently inhibited by Gd3+, but insensitive to most classical K+ channel blockers including TEA and 4-aminopyridine [6]. No previous report showed that diltiazem, a well-known Ca2+ channel blocker, inhibited TREKs. Diltiazem may be a specific blocker for TREKs although sensitivities to diltiazem should be further elucidated in other 2P domain K+ channels.

In conclusion, the present study showed novel findings that fenamate stimulate TREK and TRAAK, and diltiazem inhibit TREK, which may help to further understand functions of the mechano-gated 2P domain K+ channels.

Acknowledgements

The authors thank Bret A Hughes at University of Michigan and Hiroshi Sakurai for critical evaluation of the manuscript. This research was supported by Grant-in-Aid 11771043 and 13771018 for MT from the Ministry of Education, Japan.

Copyright information

© Springer-Verlag 2005